VOLUME 12, ISSUE 2, ARTICLE 21 Guzman, J. L., J. Caro, and B. Arroyo. 2017. Factors influencing mobility and survival of Eurasian Woodcock wintering in Spain. Avian Conservation and Ecology 12(2):21. https://doi.org/10.5751/ACE-01096-120221 Copyright © 2017 by the author(s). Published here under license by the Resilience Alliance.

Research Paper Factors influencing mobility and survival of Eurasian Woodcock wintering in Spain José Luis Guzmán 1, Jesus Caro 1 and Beatriz Arroyo 1 1Instituto de Investigación en Recursos Cinegéticos (IREC, CSIC-UCLM-JCCM)

ABSTRACT. Survival and mobility have important implications for population management for game species. These parameters are influenced by both intrinsic and extrinsic factors. We describe movements (commuting flights between diurnal refuges and nocturnal feeding places; and escape flights during cold spells) and winter survival rate of Eurasian Woodcock (Scolopax rusticola) wintering in Spain. We also evaluate factors influencing these variables, using 51 radio-tracked birds over three winters (2008/2009, 2009/2010, and 2010/2011). Commuting flight distances were estimated at 961.5 ± 1041.9 m, and variations were mainly explained by age and temperature (they decreased with lower temperatures and were lower for first-winter birds). Three cold spells occurred in 2009/2010; 80% of woodcocks monitored that winter showed escape flights, moving > 20 km, and went back to their previous wintering place when the effects of cold spells finished (about 8 days later). Of monitored woodcocks, 54.9% survived the winter. The most frequent cause of death was hunting, affecting mainly first-winter birds. Woodcock survival was lower in areas with more hunting days per week, and in Mediterranean than in Atlantic climate regions. Our results highlight the importance of monitoring survival and factors affecting it. Also, these results underline the importance of developing future studies to understand the importance of Mediterranean regions, the use of refuge places during cold spells, and hunting pressure there.

Facteurs agissant sur la mobilité et la survie de la Bécasse des bois hivernant en Espagne RÉSUMÉ. La survie et la mobilité sont des paramètres importants dans la gestion des populations pour les oiseaux gibier. Toutes deux sont influées par des facteurs tant intrinsèques qu'extrinsèques. La présente étude nous a permis de décrire les déplacements (vols quotidiens entre les refuges diurnes et les sites d'alimentation nocturne, et vols pour s'éloigner d'épisodes de froid) et la survie hivernale de la Bécasse des bois (Scolopax rusticola) hivernant en Espagne. Nous avons aussi évalué les facteurs qui agissent sur ces variables, au moyen de 51 oiseaux équipés de radio durant trois hivers (2008-2009, 2009-2010 et 2010-2011). La distance des vols quotidiens a été estimée à 961,5 ± 1041,9 m et les variations s'expliquaient principalement par l'âge et la température (la distance diminuait quand les températures étaient plus basses et était plus courte chez les oiseaux dont c'était le premier hiver). Trois épisodes de froid sont survenus en 2009-2010; 80 % des bécasses suivies au cours de cet hiver ont effectué des vols pour s'éloigner, se déplaçant > 20 km et revenant au site occupé précédemment lorsque les effets du froid ont été terminés (environ 8 jours plus tard). Parmi les bécasses suivies, 54,9 % d'entre elles ont survécu durant l'hiver. La cause la plus fréquente de mortalité était la chasse, laquelle touchait surtout les oiseaux dont c'était le premier hiver. La survie des bécasses était plus faible dans les endroits où le nombre de jour de chasse par semaine était plus élevé, ainsi que dans la région climatique méditerranéenne par rapport à celle de l'Atlantique. Nos résultats soulignent l'importance de surveiller la survie et les facteurs qui l'affectent. De plus, nous croyons qu'il serait opportun de faire d'autres recherches pour mesurer l'importance de la région méditerranéenne, l'utilisation de lieux refuges lors d'épisodes de froid, de même que la pression de chasse à ceux-ci. Key Words: daily movements; escape flights; hunting management; radio-tagging; Scolopax rusticola

INTRODUCTION example, hunting regulations may be adjusted based on Sustainable management of hunted species depends on information about when certain areas are used to avoid periods understanding its demography, but also on knowing the or areas of population vulnerability (Péron et al. 2011a, appropriate spatial scale at which decisions must be made. For Rodríguez-Teijeiro et al. 2012). highly mobile species, including migratory ones, decisions taking Mobility is known to be influenced by both intrinsic and extrinsic into account only local conditions may not be appropriate or factors. In birds, movement decisions or distances moved have effective. For example, local movement can mask the true effect been found to differ between individuals of different age or sex of harvest on observed densities in small areas (Williams et al. (e.g., Brochet et al. 2009), and are influenced by weather, habitat 2004). The study of mobility also provides information about fragmentation, pollution, or disturbance (Burger and Gochfeld ecological requirements at different times in the annual cycle 2004, Stephens et al. 2004, Gill 2007, Sauter et al. 2010). (Fauchald and Tveraa 2006, Gourlay-Larour et al. 2012), and is Understanding these relationships may also give indications therefore useful in terms of management for game species. For

Address of Correspondent: José Luis Guzmán, Ronda De Toledo 12, Ciudad Real, Ciudad Real , Spain 13005, [email protected] Avian Conservation and Ecology 12(2): 21 http://www.ace-eco.org/vol12/iss2/art21/

about whether the type and extent of movements might be related expenditure by thermoregulation in dense vegetation habitats is to variations in fitness. For example, longer flights to obtain food lower than in more open ones (Wiersma and Piersma 1994), may indicate ability to use space flexibly, therefore increasing the mainly because of the effect of wind (Bakken 1990). Therefore, chances for individual survival (Austin et al. 2006). Conversely, in the Mediterranean region, thermoregulation costs (and thus because longer flights have higher energy costs, in poor food energetic needs) under low temperatures may be lower on account conditions they could be associated with lower survival (Wikelski of greater vegetation cover. Additionally, because worm et al. 2003). availability is directly related to soil temperature (Lavelle 1988), it is also possible that the greater vegetation cover in the The Eurasian Woodcock (Scolopax rusticola), hereafter Mediterranean region results in that earthworm availability is less woodcock, is a migratory wader highly reputed as a game bird in reduced under low air temperatures there. its wintering areas of Western and Southern Europe (France, the northern parts of the Iberian Peninsula, and Italy). Studies based In this paper we assess movements and survival rate of woodcock on radio-tracking have shown that woodcock perform two types wintering in Northern Spain in sites under two climatic regions of movements during the wintering period. They make daily (Atlantic and Mediterranean) using radio-tracked birds, and movements (hereafter commuting flights) at dusk and dawn factors influencing them. We hypothesize that commuting flight between the places where they spend daytime (usually forests) and distances would be influenced by temperature, as lower other locations at night (usually grasslands). For European temperatures increase energy needs (Duriez et al. 2004b), but that wintering woodcock it has been shown that grasslands represent this relationship may be influenced by climatic region. Second, the main foraging areas, whereas forests represent more sheltered we hypothesize that survival rate would also be influenced by areas to spend the day (Cramp and Simmons 1983, Duriez et al. temperature, but that, as for commuting flight distances, the 2005a; although see Masse et al. 2013, where the opposite has relationship between temperature and survival rate may be been found for American Woodcock, Scolopax minor, in modulated by climatic region, and that survival rate would also summer). Commuting flights indicate a trade-off between using depend on hunting opportunities, i.e., number of hunting days good shelter and good feeding places, and the relative use of both allowed per week in each area (Tavecchia et al. 2002). types of places reflects the relation between energetic needs and Additionally, we describe escape flights observed, conditions predation risk (Duriez et al. 2004a). Additionally, when weather eliciting this behavior, and duration of the displacements. Finally, conditions become severe woodcock can make long-distance we discuss our results in relation to management of this species. movements (hereafter escape flights) to more temperate places where they wait until the weather conditions change and then METHODS return to their normal wintering places (Gossmann and Ferrand 2000, Péron et al. 2011a). Study species The woodcock is a forest migratory wader with a Palearctic Severe weather conditions in winter have in fact been found to distribution, breeding in temperate regions of Asia and Europe influence woodcock survival rate (Hoodless and Coulson 1994, (Cramp and Simmons 1983). It has a large breeding population Tavecchia et al. 2002, Péron et al. 2011a), which decreases when and trends appear to be stable (BirdLife International 2016), earthworms become inaccessible, as happens with cold spells in although important declines have been observed in areas like the winter (Péron et al. 2011a). On the other hand, survival rate in UK (Heward et al. 2015). Recent studies of the wintering game species may be strongly influenced by hunting pressure populations of Western Europe describe a stable tendency in itself, if mortality associated with hunting is additive instead of recent decades, although with important annual variations in compensatory (Péron 2013). Indeed, a study in France has abundance (Aebischer and Baines 2008, Ferrand et al. 2008, described spatial differences in annual survival of woodcock in Guzmán and Arroyo 2015). relation to hunting pressure (Péron et al. 2011b, 2012), with woodcock annual survival being lower in those areas with higher Study areas and years hunting pressure. Considering that breeding areas and migration We radio-tracked woodcock in winters 2008/2009, 2009/2010 and routes are similar for woodcock wintering in different areas in 2010/2011 in two provinces of Northern Spain (Fig. 1). In the France (Bauthian et al. 2007), and therefore that any mortality winter of 2008/2009, monitoring took place in three valleys of effect in winter is not likely compensated by density dependent Navarra province (Ulzama, Baztán, and Esteribar); in the winters survival later in the year, this finding supports that in certain of 2009/2010 and 2010/2011, it took place in two municipalities circumstances hunting pressure represents an additive source of of Álava province (Kuartango and Zuia; Fig. 1). The study area mortality to those found at different times in the annual cycle. was located between the sub-Atlantic and sub-Mediterranean Most previous studies about survival and mobility of woodcock climatic regions, with an irregular topography where the altitude have been developed in France, the UK, or Ireland (Wilson 1982, varies between 300 and 600 m, and landscape is dominated by Hoodless and Coulson 1994, Duriez 2003), areas dominated by forest-grassland mosaic. In the areas of Ulzama, Zuia, and Atlantic climate. Knowledge about survival and movements of Baztán Atlantic climate dominates, characterized by temperate their wintering populations in Spain is, in contrast, scarce. The weather, important rainfall throughout the year, and where only radio-tracking study so far was carried out in the Atlantic- predominant vegetation cover types are deciduous forest of beech, climate region of northern Spain (Braña et al. 2010) and there is oak, and mixed formations of these (Fagus sylvatica, Quercus no information about these topics in the Mediterranean-climate petraea, Q. robur and Q. pyrenaica) including patches of conifer region. The Mediterranean-climate region is characterized by reforestation (Pinus radiata, P. nigra and Chamaecyparis denser vegetation in the forest understory (Specht 1969). Energy lawsoniana). In contrast, in Kuartango and Esteribar Avian Conservation and Ecology 12(2): 21 http://www.ace-eco.org/vol12/iss2/art21/

Mediterranean climate dominates, with warm and drier summers, livestock blue, Nasco, Fort Atkinson, USA), following the and the predominant tree layer is composed of sclerophyllous method described in Duriez (2003). The transmitter incorporated forest (Q. ilex and Q. faginea) with dense Mediterranean scrub an activity sensor to detect if the individual was active or resting. and conifer reforestations. If an individual remained in the same geographical position without activity between two consecutive diurnal or nocturnal Fig. 1. (A) General location of the study area in Northern locations, we approached to check if it was alive or dead. A total Spain. (B) Specific locations of the radio-tracking study areas: of 61 woodcock were trapped and aged based on plumage 1, Zuia (Atlantic); 2, Kuartango (Mediterranean) (both in characteristics (Ferrand and Gossmann 2009). Four of these were Alava province); 3, Baztán (Atlantic); 4, Ulzama (Atlantic); 5, never located after tagging, in four the transmitter was incorrectly Esteribar (Mediterranean) (in Navarra province); 6, study area fitted and fell off shortly after deployment, and two died of stress during cold-spells. during the capture process. In the remaining 51 individuals (13 in 2008/2009; 17 in 2009/2010; 21 in 2010/2011), we could register either their death during the study winter (as corpses were found) or they were monitored at least until the end of February, when we were no longer able to locate them, suggesting that they had departed (prenuptial migration) and therefore survived the winter period. We identified predation as a cause of death when we found predation marks on the corpse or on a radio. In these cases, the possibility of scavenging following another source of death cannot be ruled out, for example if a hunter shot a woodcock but did not recover it, and then it was scavenged by a predator. Therefore, our predation estimates refer to “apparent predation.” Birds were identified as dead through hunting if the radio was brought back to us by local hunters. In other cases, we classified the bird as “dead for unknown reasons.” Monitoring consisted of obtaining diurnal (10:00 to 15:00) and nocturnal (20:00 to 00:00) locations of all individuals until their Study areas varied in the number of hunting days per week. death or onset of prenuptial migration. We approached woodcock Hunting was allowed seven days per week in Navarra. In Alava, by radio-triangulating locations up to 10–50 m from the bird number of hunting days allowed was usually four, but it varied during the day and 50–100 m during the night, first using a vehicle among monitored areas, because part of monitored individuals with a dipole antenna and then walking with a three-element Yagi in Zuia moved around the Natural Park of Gorbeia, where the antenna (Duriez 2003). When we located an individual, we number of hunting days per week was lower than in the rest of assessed temperature with an outdoor digital thermometer. We the areas (Fig. 1). aimed to monitor each individual at least twice a week during the day and the same frequency at night. When an individual was Weather varied among the three study winters: in 2009/2010 there found dead, we estimated time of death as the midpoint between were three cold spells in the radio-tracking study areas, one in the last time alive and when it was found dead. December, one in January, and one in February. Cold spells were defined as periods when the average temperatures were in the Commuting flight distance was calculated as the distance in lower 10 percentile for the area (in our case, below zero degrees), meters between a diurnal location and a nocturnal location, i.e., and when this situation remained more than four consecutive days the distance covered between the forest and the grasslands. The (Yagüe et al. 2006), resulting in frost and snow covering the distance was calculated with a Geographic Information System ground. In the other two winters no cold spells occurred. (QGIS 1.8.0; QGIS Development Team 2013) using consecutive day-night locations, i.e., locations in a given day and of the same Radio-tracking data individual the subsequent night. The only exception was for two Each radio-tracking season lasted approximately four months, woodcocks for which we had too few consecutive day-night starting in December and ending with the prenuptial migration locations; here, we calculated the distance between a daytime of all radio-tracked woodcock still alive at that time (late March location and the following available nighttime location, which was to early April). We captured and marked individuals during the 2–3 days later. Results excluding these day-night time pairs were first half of December, to minimize the chance of capturing no different to those presented here. For analyses, we took into migrating birds (wintering populations increased regularly until account only those individuals with three or more day-night late November, suggesting postnuptial migration until that time; location pairs (6 ± 2 pairs of locations for 34 individuals; mean Guzmán 2013). Captures were made at night in grasslands, ± SE). dazzling the bird with a flashlight while another person captures During the first cold spell observed in the 2009/2010 winter, we it with a landing net (Duriez 2003). Captured birds were fitted were unable to locate several woodcocks that were regularly with radios (TW-3 single celled tag by Biotrack) with a mass of located using the ground-based telemetry protocol, but they 12 g (< 4% of body mass). The transmitter was anchored by a returned to the study area subsequently. In the second and third simple loop forming a harness at the level of the keel and fixed cold spells that year, two light-aircraft flights were carried out with a special glue on the back of the bird (hypoallergenic Avian Conservation and Ecology 12(2): 21 http://www.ace-eco.org/vol12/iss2/art21/

after the movement of birds from the study area, to identify their location of the individuals, which were based on only one or two location. The aircraft was equipped with two unidirectional daytime and nighttime records. Furthermore, weather stations antennas and we travelled across the study area, following the give a better picture of temporal variability in conditions that method described in Gilmer et al. (1981). We monitored the match the time scale we considered, whereas the ground-level nearby area in a radius of 100 km, with emphasis in the warmer measures are more appropriate to the scale and spatial nature of and snow-free coastal areas (Fig. 1). When an individual was the commuting distance analysis. In both models, temperature detected from the aircraft, we noted the approximate geographic was standardized prior to analyses (by subtracting the mean and location using GPS and the next day a more precise location was dividing by the standard deviation) to help with model obtained from the ground following the protocol for commuting convergence. flights (see above). Escape flight distances were assessed as the We used an Information-Theoretic approach to select the best distance between the last location (day or night) before the cold models explaining variation in commuting flight distance and spell and the first location (day or night) during the cold spell. survival rate, because this approach is widely used in the wildlife Because of the small sample size (n = 15), the escape flights were and ecological literature (Burnham and Anderson 2002). We used evaluated only descriptively. the AICc differences between models with different explanatory Statistical analyses variables as criterion to select the best models, taking into account all possible combination of variables (Appendix 1). Models with We built Generalized Linear Mixed Models to assess factors less than 2 AICc points in relation to the best model were associated with variations in commuting flight distance. The Δ considered to have the same empirical support (Burnham and response variable was log-transformed to fit a normal Anderson 2002). We present the model-averaged coefficients of distribution. Individual identity was included as a random the explanatory variables included in those models, using this intercept to account for repeated measures from individuals (n = subset of models for their calculation (conditional average). 190 commuting flight distances from 34 individuals). As Variables included in those models were not collinear (Cade 2015). explanatory variables, we included temperature as a continuous We also calculated the relative variable importance, based on the variable, and climatic region (Atlantic or Mediterranean), age sum of the weights of each of the models that included this (adults or first-winter birds), and hunting opportunity (high, variable (Burnham and Anderson 2002). Analyses were carried medium, or low) as factors. Hunting opportunity was categorized out with R 3.2.4 (R Core Team 2016), using the package MuMIn in relation to the number of hunting days per week in each area (Barton 2017). Means are presented ± SE. as “high” if hunting occurred seven days a week, “medium” if hunting occurred four days per week, or “low” when there were less than four hunting days per week. Temperature for this analysis RESULTS was calculated as the average of the diurnal and nocturnal temperature records obtained for each individual in the day/night Commuting flights location considered for each commuting flight. We also included Commuting flight distances ranged between 10.0 m and 5135.1 in the model the interaction between region and temperature. m (mean 961.5 ± 1041.9 m, n = 190). The most important variables explaining variation in commuting flight distance were age group For analyses of winter survival rate, we built Generalized Linear and temperature, variables that appeared in all the selected models Mixed Models with the response variable “weekly survival” fitted (Tables 1 and 2). Commuting flights were shorter in first-winter to a binomial distribution (with a logit link), and following the birds than adults (1380.48 ± 1260.11 m for adults, n = 94; 702.75 basic assumptions of telemetry-based survival estimation ± 702.75 m for first-winter birds, n = 75), and were shorter under (Murray 2006). This variable was scored as 1 if an individual had lower temperatures (Table 2). remained alive in a particular week, and 0 if it had died during that week. Individual identity was included as a random intercept Table 1. Results of models explaining woodcock (Scolopax to account for repeated measures of individuals (n = 551 weekly rusticola) commuting flight distance. The table contains the full registers from 51 individuals). As above, explanatory variables model and selected models (models with AIC less than 2 points included temperature, climatic region, age, hunting opportunity, Δ from the best one). K = number of parameters. W = weight of and the interaction between temperature and region. We did not the model. consider commuting flight distance as an explanatory variable because sample size did not allow it: woodcock that died relatively early in the winter were not monitored sufficiently long to obtain Models K AICc ΔAICc W enough night-day consecutive locations. For this analysis, Selected models temperature was calculated as the average minimum temperature ~ Temperature + Age + Hunting 7 625.6 0 0.46 of all days of a given week from the registers of the nearest climatic ~ Temperature + Age + Region 6 626.7 1.1 0.27 ~ Temperature + Age + Hunting + Region 8 626.8 1.1 0.27 station to the study area (Pamplona climatic station in 2008/2009, Full model and Vitoria climatic station in 2009/2010 and 2010/2011; https:// ~ Temperature + Age + Hunting + Region + 9 629.3 3.7 - www.tutiempo.net/clima). The approximate distances to the Temperature*Region climate stations from the field sites were: Pamplona-Esteribar, 15 km; Pamplona-Ulzama, 20 km; Pamplona-Baztán, 30 km; The three best models also contained either hunting opportunity, Vitoria-Zuia, 15 km; Vitoria-Kuartango, 20 km. We considered region, or both as explanatory variables (Table 1). Flight distances this to be a more reliable indication of the temperature appeared to be longer in Atlantic than Mediterranean regions, experienced during that week than our records during the weekly and under low more than medium and high hunting disturbance, Avian Conservation and Ecology 12(2): 21 http://www.ace-eco.org/vol12/iss2/art21/

although the parameter coefficients for some of the categories occurring between 21 February and 3 April. These birds were within those variables were unreliable (Table 2). The interaction inferred to have undergone prenuptial migration and were between temperature and region did not appear in any of the best therefore considered to have survived the winter. Among those models. that did not survive, 47.8% of individuals were hunted (13.0% of adults and 34.8% of first-winter birds), 21.7% were killed by Table 2. Model-averaged parameter coefficients of variables predators (17.4% of adults and 4.3% of first-winter birds), and explaining woodcock (Scolopax rusticola) commuting flight in 30.4% it was not possible to determine the cause of death (17.4% distance from selected models. CI = 95% confidence intervals, and 13.0% of adults and first-winter birds; Table 4). Woodcock RVI = relative variable importance. that survived (n = 28) remained in the winter area an average of 105.9 ± 11.6 days from the beginning of December, (109.0 ± 11.6 days for adults, n = 19; and 110.1 ± 9.1 days for first-winter birds, Explanatory variables Coefficient SE CI CI RVI n = 11). (2.5%) (97.5%) Intercept 7.41 0.40 6.60 8.14 - Age (first-winter birds) -1.38 0.40 -2.16 -0.59 1 Table 4. Number of radio-tracked individuals in relation to fate Temperature -0.98 0.53 -0.39 -0.07 1 in the different seasons and age groups. Hunting (medium) -0.95 0.61 -1.90 -0.20 0.73 Hunting (high) -0.23 0.08 -2.14 0.22 0.73 Region (Mediterranean) -0.52 0.60 -1.60 0.03 0.54 Fate Winter Age Migrated Hunted Depredated Dead (Unknown reason) Escape flights 2008/09 Adult 1 1 1 3 Of the 20 woodcock monitored in 2009/2010, only four did not First-winter 2 4 0 1 make escape flights during cold spells. All others moved in at least 2009/10 Adult 11 1 2 0 one of the cold spells from the study area to other valleys First-winter 1 0 0 2 (hereafter refuge areas; Fig. 1), characterized from being at lower 2010/11 Adult 5 1 1 1 elevation (200 m above sea level instead of 600 m on average for First-winter 8 4 1 0 Total 28 11 5 7 the study areas), and where ground was only partially covered with snow. The distance that woodcock moved during escape flights between the usual wintering areas and refuge areas was on The most important variable explaining variation in survival rate average 23.4 ± 6.0 km (Table 3). The birds remained in the refuge was hunting opportunity, appearing in all of the best models areas an average of eight days and went back to within 77 m of (Table 5). Winter survival rate was increasingly higher in areas their previous wintering place (Table 3). Only one bird did not go with lower hunting opportunities (Table 6 and 7). Age, region, back to its original wintering area, and remained during the rest and temperature also appeared each in one of the three best of the winter in the refuge area. models (Table 5): winter survival was higher in Mediterranean- than Atlantic-climate region, in adults than in first-winter birds, and increased with increasing temperature (Table 6 and 7). Table 3. Parameters describing the displacements occurred during However, the relative importance of these variables was much the three cold spells of 2009/2010. Figures in parentheses show lower and parameter coefficients included zero when taking into sample size used in each case (although the occurrence or not of account their confidence intervals (Table 6). The interaction escape flights was monitored in all individuals, we were not able between temperature and region did not appear in any of the best to identify the other parameters for all of these individuals). models.

Cold Frequency of Distance Time in Fidelity of spell escape flights covered refuge return (m) Table 5. Results of models explaining winter woodcock (Scolopax (%)† (km)‡ (days) rusticola) survival. The table contains the full model and selected December§ 53.3 (15) 27.3 ± 3.3 (3) - - models (models with ΔAIC less than 2 points from the best one). January 64.3 (14) 22.2 ± 6.6 (4) 8.2 ± 2.2 (8) 76.1 ± 4.4 (3) K = number of parameters. W = weight of the model. February 64.3 (14) 19.6 ± 6.4 (3) 7.2 ± 1.5 (9) 77.7 ± 20.8 (3) † Proportion of monitored individuals in 2009/2010 that did escape Models K AICc ΔAICc W flights. ‡ Selected models Distance between the location before the cold spell and the first ~ Hunting 4 187.8 0.0 0.41 location after displacement. § ~ Hunting + Region 5 188.6 0.9 0.26 In the cold spell of December, it was impossible to describe the time ~ Hunting + Temperature 5 189.6 1.8 0.16 that they spent in the refuge or the fidelity when they came back to their ~ Hunting + Age 5 189.6 1.8 0.16 usual wintering places. Full model ~ Hunting + Temperature + Age + Region + 8 194.5 6.7 - Temperature*Region Survival rate During the radio-tracking period, 54.9% of tagged woodcock (60.7% of adults and 47.8 % of first-winter birds) were monitored until the end of the winter season (Table 4), with the last location Avian Conservation and Ecology 12(2): 21 http://www.ace-eco.org/vol12/iss2/art21/

travelling to better foraging places is compensated by lower Table 6. Model-averaged coefficients of the variables explaining thermoregulatory costs under higher temperatures. In our case, winter woodcock (Scolopax rusticola) survival. CI = 95% we found a negative correlation between temperature and confidence intervals, RVI = relative variable importance. commuting flight distance. According to Duriez et al. (2004a), who found similar results for Eurasian Woodcock in France, this Explanatory variables Coefficient SE CI CI RVI pattern could be the result of a compensatory mechanism in (2.5%) (97.5%) response to higher energy needs for thermoregulation in cold Intercept 4.15 0.72 2.72 5.54 - temperatures. Braña et al. (2010) found in Atlantic Spain a Hunting opportunity -0.71 0.88 -2.49 1.01 1 positive correlation between temperature and the time initiating (medium) commuting flights after sunset, with flights starting earlier in Hunting opportunity -2.10 0.86 -3.82 -0.42 1 colder weather. Overall, these results suggest that when (high) temperature falls, increasing energy expenditure by thermoregulation, Region (Mediterranean) 0.75 0.71 -0.63 2.14 0.26 Temperature -0.11 0.23 -0.56 0.35 0.16 woodcock invest more in foraging flights, with higher activity and Age (first-winter birds) -0.25 0.54 -1.31 0.81 0.16 greater displacements, probably in an attempt to optimize feeding (Charnov 1976). The extreme case would be represented by the responses to cold spells, when movement distances were an order of magnitude higher than usual. In France, responses to cold spells of different intensity have been studied, and a threshold in Table 7. Winter survival probabilities (Swinter = Sweek ^ Nweeks, relation to the magnitude of the cold spell has been found. Thus, Powell 2007) in relation to hunting opportunity, region, and age when the cold spell is not very intense, few individuals change group. Weekly survival rates estimated from model averaged their usual behavior and displacements are shorter, whereas when coefficients in Table 6. We estimated the length of winter as 16 it exceeds a certain level, a higher number of individuals show weeks (from December to March). escape behaviors and the magnitude of the displacements is higher as well (Péron et al. 2011a). This seems consistent with the results Atlantic Mediterranean of the present paper, in which the largest displacements occurred Hunting Low Medium High Low Medium High in the winter 2009/2010 under particularly harsh weather opportunity conditions, and did not influence equally all individuals: although Adults 0.78 0.60 0.14 0.89 0.79 0.39 some individuals moved to refuge areas, others faced the harsh First-winter birds 0.73 0.52 0.09 0.86 0.74 0.30 conditions in their usual wintering places. Therefore, this compensatory mechanism appears to act at different levels in a gradient from low to high intensity, from daily travels to feed in the grasslands near the forest (commuting flights), to the great DISCUSSION journeys to more temperate zones (escape flights), conditioning Our data support the hypothesis that commuting flight distances woodcock’s winter mobility at different temporal and spatial are influenced by temperature, and that winter survival rate was scales. influenced by hunting. On the other hand, we did not find support for the hypothesis that the relationship between temperature and In contrast to our hypothesis, we did not find differences among either mobility or survival rate was influenced by climatic region regions in commuting flight distances in relation to temperature. (although results suggested that survival was higher in This may indicate that our initial hypothesis (that in Mediterranean than Atlantic regions). Mediterranean-climate region, higher shrub vegetation cover modifies thermoregulation costs for woodcock or worm Commuting and escape flights accessibility under low air temperatures) is not supported. In fact, Our results showed that age and temperature were important habitat utilization by predators can reflect availability of resources factors related to the length of commuting flights for woodcock (Davoren et al. 2003), so our results may indicate that both wintering in Spain. In relation to age, we found that adults moved climatic regions in fact have similar food resources, e.g., number on average longer distances than first-winter birds. Similar results and distance of grasslands where to feed on earthworms. Further were found for woodcock in Ireland (Wilson 1982) and for other studies should evaluate whether abundance and distribution of migratory birds, which has been interpreted as an indication that grasslands and forest patches is an important factor for winter the location of the most appropriate (even if further) foraging ecology of woodcock, and if there are differences in food areas is learned through experience (Riotte-Lambert and resources, spatial distribution of forest and grassland areas, or Weimerskirch 2013). Additionally, adult birds may store more fat soil temperature according to climate regions. than first-winter birds (Duriez et al. 2004b), which may result in them being better able to meet the increased demands from Winter survival rate performing longer flights for food. Our results thus support that The most frequently identified cause of death of our tracked first-winter birds may be less efficient at finding good foraging woodcock was hunting, affecting mainly first-winter birds. areas and less proficient at optimizing the time spent feeding Number of woodcock killed by predation were only half in (Duriez et al. 2004a). relation to hunted birds. This difference may even be underestimated, because the four woodcock that were never A study on American Woodcock showed that woodcock moved located after tagging disappeared in areas with high hunting increasingly further distances with higher temperatures (Doherty opportunity, so it is possible that they were hunted by hunters et al. 2010), which was interpreted as meaning that the cost of Avian Conservation and Ecology 12(2): 21 http://www.ace-eco.org/vol12/iss2/art21/

who did not return the transmitter, but this was impossible to winter mortality of woodcock wintering in Spain may be high in evaluate. relation to other areas, at least in certain conditions. Indeed, winter survival estimated for the areas with high hunting Furthermore, our study showed that survival rate was lower in opportunities, if confirmed in further studies, could reflect that areas with higher hunting opportunity. Survival rate was populations there are unsustainable without annual recruitment markedly low when it was possible to hunt all days of the week of first-winter birds (Perón et al. 2011b). Right now, populations (from 0.13 to 0.20 depending of region and age group). It is seem stable (BirdLife International 2016), but it appears strongly important to underline that because of our work design, the necessary to continue monitoring both hunting pressure and variable hunting opportunity was somewhat collinear with winter survival rates to identify which places could act as a sink and to because the category “high” in hunting opportunity only occurred be able to adjust management if necessary to allow population in one of the study winters. Therefore, it is not possible to sustainability (Péron et al. 2011b, 2012). determine if part of the variance was explained by differences among study years in relation to factors unmonitored in the study. Our results also suggest that survival is higher in Mediterranean However, differences in survival rates between all categories of rather than Atlantic climate region, supporting the hypothesis hunting opportunity followed a gradual pattern, and survival rate that in Mediterranean regions, thermoregulation costs would be was also 11–28% lower in areas with medium than low hunting lower or worm availability be higher, and therefore that these opportunities (calculated from Table 7), which were sampled in regions could be important for woodcock. However, the two different winters, so hunting opportunity appears to have an confidence interval for the effect size of this variable included effect on survival independently of interannual differences. zero, so these results should be confirmed. It may be particularly useful if future studies on winter survival of woodcock in Spain More hunting opportunities, i.e., a higher number of days when consider this variable. hunting is allowed, do not necessarily indicate higher hunting pressure, and even the number of hunting days may be reactively In contrast to our hypothesis, we did not find a marked effect of reduced in areas where the hunting pressure is too high. However, temperature on winter survival rate, in either of the two climatic there are some clues that indicate that, in our case, the positive regions. With low temperatures, nonhibernating endotherms are relationship between this parameter and hunting pressure is likely. confronted with increasing energy requirements to enhance In Guipuzkoa (a province in Northern Spain close to our study thermogenesis while food is a scarce resource (Boos et al. 2007, sites), woodcock hunting is also allowed every day of the week. Deville et al. 2014). In many bird species it has thus been shown In that area, there exists an administrative mechanism whereby that winter conditions such as low temperatures or consecutive hunters have to inform the authorities about their daily hunting frosts influence survival (e.g., Deville et al. 2014, Johnston et al. activity (including hunting bags). Results from that area show 2016). In France, a negative relationship between winter how hunting actually occurs with high intensity (in terms of temperature and woodcock survival was also found, as well as number of hunting sessions) each weekday, and that spatially escape behavior below a certain temperature threshold (Tavecchia there exists a positive correlation between the total hunting et al. 2002, Péron et al. 2011a), presumably because dealing with sessions per area and total woodcock hunted (n = 11 areas, r = such conditions would be riskier than moving away from their 0.96; based on information in García and Romero 2015, usual wintering areas (Duriez 2003). The fact that we did not find unpublished data). This suggests that increasing the number of a significant relationship between temperature and survival rate, days when hunting is possible probably increases the total number may reflect that the measure of temperature used for this analyses of hunting sessions in an area, and thus hunting pressure. was not sufficiently precise because of the distance between the climatic stations and the areas used by woodcock. However, and Our results coincide with other woodcock studies that indicate alternatively, it may also be a consequence of the closeness of that hunting mortality is an important factor explaining survival suitable refuge areas found in our study, which may provide an rate. These studies also suggest that hunting is an additive escape for woodcock to compensate the negative effects of low mortality factor, because annual survival rates for woodcock temperatures. Most woodcock showed escape flights during the wintering in areas without hunting (Duriez et al. 2005b, Aradis cold spells, and although the refuge areas were used only for et al. 2008) was higher than annual survival rate for woodcock in limited amounts of time, they could be very important for the areas with hunting activity (Hoodless and Coulson 1994, birds during harsh conditions. This underlines the importance of Tavecchia et al. 2002, Bauthian et al. 2006; see also Bruggink et developing future studies to understand how they are used, and al. 2013 for American Woodcock). According to that, and given whether or not woodcock density or harvest pressure is greater that woodcock winter survival rates in this study varied in relation there during cold spells. to hunting opportunity, this may indicate source-sink population dynamics among areas with different hunting pressure, as described in France (Péron et al. 2011b and 2012). Responses to this article can be read online at: The winter survival rate found in our study (0.55, n = 51) was http://www.ace-eco.org/issues/responses.php/1096 lower than that reported for France for both adults and first- winter birds (0.77 and 0.59, respectively; Tavecchia et al. 2002), and also lower than the annual survival rate reported for adults Acknowledgments: in the UK (0.58; Hoodless and Coulson 1994). It was only higher than that reported for first-winter birds in the UK (0.47; Hoodless This work was partially funded by the Committee of Hunting and and Coulson 1994). This is a first estimate, and may be influenced Continental Fishing, under the Research Project: 200430E471. We by our limited sample size. However, this may also suggest that are grateful to all hunters who participated in the capturing and Avian Conservation and Ecology 12(2): 21 http://www.ace-eco.org/vol12/iss2/art21/

marking of individuals, especially Ibon Telletxea and Rubén Ibanez, Burnham, K. P., and D. R. Anderson. 2002. Model selection and and two anonymous referees for constructive and useful comments multimodel inference: a practical information-theoretic approach. that improved the manuscript. Iñigo Mendiola gave us useful Second edition. Springer-Verlag, New York, New York, USA. information about hunting in Guipuzkoa province. http://dx.doi.org/10.1007/b97636 Cade, B. S. 2015. Model averaging and muddled multimodel LITERATURE CITED inferences. Ecology 96(9):2370-2382. http://dx.doi.org/10.1890/14-1639.1 Aebischer, N. J., and D. Baines. 2008. Monitoring gamebird Charnov, E. L. 1976. Optimal foraging, the marginal value abundance and productivity in the UK: the GWCT long-term theorem. Theoretical Population Biology 9(2):129-136. http://dx. datasets. Revista Catalana d’Ornitologia 24:30-43. doi.org/10.1016/0040-5809(76)90040-X Aradis, A., M. W. Miller, G. Landucci, P. Ruda, S. Taddei, and F. Cramp, S., and K. E. L. Simmons. 1983. Handbook of the Birds Spina. 2008. Winter survival of Eurasian Woodcock Scolopax of Europe, the Middle East and North Africa, III Waders to Gulls. rusticola in central Italy. Wildlife Biology 14:36-43. http://dx.doi. Oxford University Press, Oxford, UK. org/10.2981/0909-6396(2008)14[36:WSOEWS]2.0.CO;2 Davoren, G. K., W. A. Montevecchi, and J. T. Anderson. 2003. Austin, D., W. D. Bowen, J. I. McMillan, and S. J. Iverson. 2006. Distributional patterns of a marine bird and its prey: habitat Linking movement, diving, and habitat to foraging success in a selection based on prey and conspecific behaviour. Marine large marine predator. Ecology 87:3095-3108. http://dx.doi. Ecology Progress Series 256:229-242. http://dx.doi.org/10.3354/ org/10.1890/0012-9658(2006)87[3095:LMDAHT]2.0.CO;2 meps256229 Bakken, G. S. 1990. Estimating the effect of wind on avian Deville, A.-S., S. Labaude, J.-P. Robin, A. Béchet, M. Gauthier- metabolic rate with standar operative temperature. Auk Clerc, W. Porter, M. Fitzpatrick, P. Mathewson, and D. Grémillet. 107:587-594. 2014. Impacts of extreme climatic events on the energetics of long- lived vertebrates: the case of the Greater Flamingo facing cold Barton, K. 2017. MuMIn: multi-model inference. R package spells in the Camargue. Journal of Experimental Biology version 1.40.0. R Foundation for Statistical Computing, Vienna, 217:3700-3707. http://dx.doi.org/10.1242/jeb.106344 Austria. [online] URL: https://cran.r-project.org/web/packages/ MuMIn/ Doherty, K. E., D. E. Andersen, J. Meunier, E. Oppelt, R. S. Lutz, and J. G. Bruggink. 2010. Foraging location quality as a predictor Bauthian, I., F. Gossmann, Y. Ferrand, and R. Julliard. 2007. of fidelity to a diurnal site for adult female American Woodcock Quantifying the origin of woodcock wintering in France. Journal Scolopax minor. Wildlife Biology 16:379-388. http://dx.doi. of Wildlife Management 71(3):701-705. http://dx.doi. org/10.2981/09-100 org/10.2193/2004-354 Duriez, O. 2003. Individual wintering strategies in the Eurasian Bauthian, I., I. Illjinsky, S. Fokin, R. Julliard, F. Gossmann, and Woodcock Scolopax rusticola: energetic trade-offs for habitat Y. Ferrand. 2006. Survival rates of Russian woodcocks. selection. Dissertation,.Université de Paris VI, Paris, France. International Wader Studies 13:62-65. Duriez, O., C. Eraud, C. Barbraud, and Y. Ferrand. 2005b. Factors BirdLife International. 2016. Scolopax rusticola. The IUCN Red affecting population dynamics of Eurasian Woodcocks wintering List of Threatened Species, Cambridge, UK. [online] URL: http:// in France: assessing the efficiency of a hunting-free reserve. dx.doi.org/10.2305/IUCN.UK.2016-3.RLTS.T22693052A86627978. Biological Conservation 122:89-97. http://dx.doi.org/10.1016/j. en biocon.2004.07.002 Boos, M., T. Zorn, G. Delacour, and J. P. Robin. 2007. Weather Duriez, O., Y. Ferrand, F. Binet, E. Corda, F. Gossmann, and H. and body condition in wintering mallards Anas platyrhynchos. Fritz. 2005a. Habitat selection of the Eurasian Woodcock in Bird Study 54:154-159. http://dx.doi.org/10.1080/00063650709461470 winter in relation to earthworms availability. Biological Braña, F., L. Prieto, and P. González-Quirós. 2010. Habitat Conservation 122(3):479-490. http://dx.doi.org/10.1016/j. change and timing of dusk flight in the Eurasian Woodcock: a biocon.2004.08.011 trade-off between feeding and predator avoidance? Annales Duriez, O., H. Fritz, F. Binet, Y. Tremblay, and Y. Ferrand. 2004a. Zoologici Fennici 47:206-214. http://dx.doi.org/10.5735/086.047.0305 Individual activity in wintering Eurasian Woodcocks: starvation Brochet, A.-L., M. Guillemain, C. Lebarbenchon, G. Simon, H. versus predation risk trade-off? Animal Behaviour 69:39-49. Fritz, A. J. Green, F. Renaud, F. Thomas, and M. Gauthier-Clerc. http://dx.doi.org/10.1016/j.anbehav.2004.04.009 2009. The potential distance of highly pathogenic avian influenza Duriez, O., L. Pastout Lucchini, M. Boos, O. Chastel, H. Fritz, virus dispersal by Mallard, Common Teal and Eurasian Pochard. Y. Ferrand, and J. Clobert. 2004b. Low levels of energy EcoHealth 6:449-457. http://dx.doi.org/10.1007/s10393-010-0275-4 expenditure in a nocturnal, forest-dwelling wader, the Eurasian Bruggink, J. G., E. J. Oppelt, K. E. Doherty, D. E. Andersen, J. Woodcock Scolopax rusticola. Ardea 92:31-42. Meunier, and R. S. Lutz. 2013. Fall survival of American Fauchald, P., and T. Tveraa. 2006. Hierarchical patch dynamics woodcock in the western Great Lakes Region. Journal of Wildlife and animal movement pattern. Oecologia 149:383-395. http://dx. Management 77(5) 1021-1030. http://dx.doi.org/10.1002/jwmg.547 doi.org/10.1007/s00442-006-0463-7 Burger, J., and M. Gochfeld. 2004. Marine birds as sentinels of Ferrand, Y., and F. Gossmann. 2009. La bécasse des bois. Histoire environmental pollution. EcoHealth 1:263-274. http://dx.doi. naturelle. Effet de lisière-éditeur, Saint-Lucien, France. org/10.1007/s10393-004-0096-4 Avian Conservation and Ecology 12(2): 21 http://www.ace-eco.org/vol12/iss2/art21/

Ferrand, Y., F. Gossmann, C. Bastat, and M. Guénézan. 2008. Péron, G. 2013. Compensation and additivity of anthropogenic Monitoring of the wintering and breeding Woodcock populations mortality: life-history effects and review of methods. Journal of in France. Revista Catalana d'Ornitologia 24:44-52. Animal Ecology 82:408-417. http://dx.doi.org/10.1111/1365-2656.12014 Gill, J. A. 2007. Approaches to measuring the effects of human Péron, G., Y. Ferrand, R. Choquet, R. Pradel, F. Gossmann, C. disturbance on birds. Ibis 149(Suppl1):9-14. http://dx.doi. Bastat, M. Guénézan, I. Bauthian, R. Julliard, and O. Gimenez. org/10.1111/j.1474-919X.2007.00642.x 2012. Spatial heterogeneity in mortality and its impact on the population dynamics of Eurasian Woodcocks. Population Gilmer, D. S., L. M. Cowardin, R. L. Duval, L. M. Mechlin, and Ecology 54: 305-312. http://dx.doi.org/10.1007/s10144-012-0309-6 C. W. Shaiffer. 1981. Procedures for the use of aircraft in wildlife biotelemetry studies. U.S. Fish and Wildlife Service Resource Péron, G., Y. Ferrand, F. Gossmann, C. Bastat, M. Guénézan, Publication, Washington, D.C., USA. and O. Gimenez. 2011a. Escape migration decisions in Eurasian Woodcocks: insights from survival analyses using large-scale Gossmann, F., and Y. Ferrand. 2000. Impact of the 1996-97 cold recovery data. Behavioral Ecology and Sociobiology 65:1949-1955. spell on Woodcock in France based on ring recoveries. Pages 37-39 http://dx.doi.org/10.1007/s00265-011-1204-4 in H. Kalchreuter, editor. Fifth European Woodcock and Snipe Workshop: Proceedings of an International Symposium of the Péron, G., Y. Ferrand, F. Gossmann, C. Bastat, M. Guenezan, Wetlands International Woodcock and Snipe Specialist Group, and O. Gimenez. 2011b. Nonparametric spatial regression of Czempin, Poland. Wetlands International, Wageningen, The survival probability: visualization of population sinks in Eurasian Netherlands. Woodcock. Ecology 92:1672-1679. http://dx.doi.org/10.1890/10-2224.1 Gourlay-Larour, M.-L., V. Schricke, C. Sorin, M. L’Hostis, and Powell, L. A. 2007. Approximating variance of demographic A. Caizergues. 2012. Movements of wintering diving ducks: new parameters using the delta method: a reference for avian insights from nasal saddled individuals. Bird Study 59:266-278. biologists. Condor 109(4):949-954. http://dx.doi.org/10.1650/0010-5422 http://dx.doi.org/10.1080/00063657.2012.697867 (2007)109[949:AVODPU]2.0.CO;2 Guzmán, J. L. 2013. Factores que modulan la abundancia QGIS Development Team. 2013. QGIS Geographic Information poblacional de la becada (Scolopax rusticola): implicaciones para System. Open Source Geospatial Foundation Project. [online] su gestión y conservación. Dissertation. Universidad de Castilla URL: http://qgis.osgeo.org La Mancha, Spain. R Core Team. 2016. R: A language and environment for statistical Guzmán, J. L., and B. Arroyo. 2015. Predicting winter abundance computing. R Foundation for Statistical Computing, Vienna, of woodcock Scolopax rusticola using weather data: implications Austria. [online] URL: https://www.R-project.org/ for hunting management. European Journal of Wildlife Research Riotte-Lambert, L., and H. Weimerskirch. 2013. Do naive 61:467-474. http://dx.doi.org/10.1007/s10344-015-0918-4 juvenile seabirds forage differently from adults? Proceedings of Heward, C. J., A. N. Hoodless, G. J. Conway, N. J. Aebischer, S. the Royal Society B-Biological Sciences 280:20131434. http://dx. Gillings, and R. J. Fuller. 2015. Current status and recent trend doi.org/10.1098/rspb.2013.1434 of the Eurasian Woodcock Scolopax rusticola as a breeding bird Rodríguez-Teijeiro, J. D., F. Sardà-Palomera, and M. Puigcerver. in Britain. Bird Study 62:535-551. http://dx.doi.org/10.1080/000­ 2012. Post-breeding movements and migration patterns of 63657.2015.1092497 western populations of Common Quail (Coturnix coturnix): from Hoodless, A. N., and J. C. Coulson. 1994. Survival rates and knowledge to hunting management. Animal Biodiversity and movements of British and continental Woodcock Scolopax Conservation 35:333-342. rusticola in the British Isles. Bird Study 41:48-60. http://dx.doi. Sauter, A., F. Korner-Nievergelt, and L. Jenni. 2010. Evidence of org/10.1080/00063659409477197 climate change effects on within-winter movements of European Johnston, A., R. A. Robinson, G. Gargallo, R. Julliard, H. Van Mallards Anas platyrhynchos. Ibis 152:600-609. http://dx.doi. Der Jeugd, and S. R. Baillie. 2016. Survival of Afro-Palaearctic org/10.1111/j.1474-919X.2010.01028.x passerine migrants in western Europe and the impacts of seasonal Specht, R. L. 1969. A comparison of the sclerophyllous vegetation weather variables. Ibis 158:465-480. http://dx.doi.org/10.1111/ characteristic of Mediterranean type climates in France, ibi.12366 California, and southern Australia. I. Structure, morphology, and Lavelle, P. 1988. Earthworm activities and the soil system. Biology succession. Australian Journal of Botany 17(2):277-292. http://dx. and Fertility of Soils 6:237-251. http://dx.doi.org/10.1007/ doi.org/10.1071/BT9690277 BF00260820 Stephens, S. E., D. N. Koons, J. J. Rotella, and D. W. Willey. 2004. Masse, R. J., B. C. Tefft, J. A. Amador, and S. R. McWilliams. Effects of habitat fragmentation on avian nesting success: a review 2013. Why woodcock commute: testing the foraging-benefit and of the evidence at multiple spatial scales. Biological Conservation predation-risk hypotheses. Behavioral Ecology 24(6):1348-1355. 115:101-110. http://dx.doi.org/10.1016/S0006-3207(03)00098-3 http://dx.doi.org/10.1093/beheco/art073 Tavecchia, G., R. Pradel, F. Gossmann, C. Bastat, Y. Ferrand, Murray, D. L. 2006. On improving telemetry-based survival and J. D. Lebreton. 2002. Temporal variation in annual survival estimation. Journal of Wildlife Management 70(6):1530-1543. probability of the Eurasian Woodcock Scolopax rusticola http://dx.doi.org/10.2193/0022-541X(2006)70[1530:OITSE]2.0.CO;2 wintering in France. Wildlife Biology 8:21-30. Avian Conservation and Ecology 12(2): 21 http://www.ace-eco.org/vol12/iss2/art21/

Wiersma, P., and T. Piersma. 1994. Effects of microhabitat, flocking, climate and migratory goal on energy expenditure in the annual cycle of red knots. Condor 96:257-279. http://dx.doi. org/10.2307/1369313 Wikelski, M., E. M. Tarlow, A. Raim, R. H. Diehl, R. P. Larkin, and G. H. Visser. 2003. Avian metabolism: costs of migration in free-flying songbirds. Nature 423(6941):704-704. http://dx.doi. org/10.1038/423704a Williams, C. K., R. S. Lutz, R. D. Applegate, and W. Palmer. 2004. Winter survival and additive harvest in Northern Bobwhite coveys in Kansas. Journal of Wildlife Management 68(1):94-100. http:// dx.doi.org/10.2193/0022-541X(2004)068[0094:WSAAHI]2.0.CO;2 Wilson, H. J. 1982. Movements, home ranges, and habitat use of wintering woodcocks in Ireland. U.S. Fish and Wildlife Service Wildlife Research Report 14:168-178. Yagüe, C., M. Martija, J. Torres, A. Maldonado, and E. Zurita. 2006. Análisis estadístico de las olas de calor y frío en España. XXIX Jornadas Científicas de la AME 20-26. [online] URL: http:// pendientedemigracion.ucm.es/info/Geofis/g-micromet_varclima/ articulos/Yague_et_alAME2006.pdf

Editor-in-Chief: Ryan Norris Subject Editor: Erik Blomberg Appendix 1. AIC of all possible models resulting from the combinations of the variables included in the model selection procedure.

Models K AICc ΔAICc W ~ Hunting 4 187.8 0.0 0.26 ~ Hunting + Region 5 188.6 0.9 0.17 ~ Hunting + Temperature 5 189.6 1.8 0.11 ~ Hunting + Age 5 189.6 1.8 0.11 ~ Hunting + Region + Temperature 6 190.5 2.7 0.07 ~ Hunting + Age + Region 6 190.7 2.9 0.06 ~ Hunting + Age + Temperature 6 191.4 3.7 0.04 ~ Null model 2 191.7 3.9 0.04 ~ Hunting + Region + Temperature + Region * Temperature 7 192.4 4.7 0.03 ~ Hunting + Age + Region + Temperature 7 192.5 4.7 0.03 ~ Region 3 192.9 5.1 0.02 ~ Temperature 3 193 5.2 0.02 ~ Age 3 193.4 5.6 0.02 ~ Region + Temperature 4 194.3 6.5 0.01 ~ Hunting + Age + Region +Temperature+Region * Temperature 8 194.5 6.7 0.01 ~ Age + Temperature 4 194.8 7.0 0.01 ~ Age + Region 4 194.8 7.1 0.01 ~ Region + Temperature + Region * Temperature 5 196.2 8.5 0.00 ~ Age + Region + Temperature 5 196.2 8.5 0.00 ~ Age + Region + Temperature + Region * Temperature 6 198.2 10.4 0.00